Fact · Speech recognition
The SALMONN authors say their model lets a large language model process and understand general audio inputs directly, with competitive performance on speech tasks such as automatic speech recognition.
Checked against its source on October 17, 2026
In the source's own words
“SALMONN enables the LLM to directly process and understand general audio inputs and achieve competitive performances on a number of speech and audio tasks used in training, such as automatic speech recognition and translation”
SALMONN: Towards Generic Hearing Abilities for Large Language Models, Changli Tang et al., arXiv (ICLR 2024), 2023, Abstract.
Open the original- What it measures
- SALMONN connects a speech encoder and an audio encoder to a pre-trained language model. It is an early example of the audio-language model design that later speech systems adopted.
- What it does not show
- The phrase competitive performance is the authors' own assessment on their benchmarks, not a ranking against the best dedicated recognizers.